Timeseries analysis top edges

Introduction

Here is a study of timeseries to evaluate the data we have and be able to discretize the time.

Index:

• Speed.
• Intersession time.
• Standard deviation.

---

With this first approach we can see the almost total absence of points on the weekends. Also, very few points at night in all edges. Most points are under 50km/h.

Timeseries

Speed Boxplot.

Median is under 5km/h for every hour and percentile 75 under 30km/h. At 13h is the time with less difference and lowest speed records.

Boxplot intersession time

The intersession time median it is not even under 60min for every edge. Frei’s st. shows an intersession time way smaller than the rest.

All speed timeseries per way_id

Avg speed by session and then time.

[[1]]

[[2]]

[[3]]

[[4]]

[[5]]

[[6]]

[[7]]

[[8]]

[[9]]

Hour

Week day

Day

Average of speed of every point.

Number of sessions

• We can see usually a incremental of number of sessions along the week. Being thursdays and fridays the one with most sessions. *The drop in day 25th it is because holiday.

[[1]]

[[2]]

[[3]]

[[4]]

[[5]]

[[6]]

[[7]]

[[8]]

[[9]]

NA

Average speed

[[1]]

[[2]]

[[3]]

[[4]]

[[5]]

[[6]]

[[7]]

[[8]]

[[9]]

NA

Intersession time

Error in df_intersession_april16pt %>% filter(way_id == df_osm_edge_ids[index,  : 
  could not find function "%>%"

Density lines average speed all together

Standard deviation

By hour

By day and hour

By session

–>